Amy Bray

ContingentAI

Workforce analytics + Leave compliance for the workers everyone forgets.

Designed from $60M+ in contingent workforce operations, ContingentAI is an AI-assisted prototype that tracks FMLA and PFML eligibility for W-2 contractors—the compliance gap most companies don't know they have.

📖 Read the Design Story View Features

The Challenge

After managing contractor payroll at Upwork ($80M+) and Velocity Global, I saw the same problem everywhere:

  • Long-term W-2 contractors qualify for FMLA after 1,250 hours—but companies don't track this
  • State PFML programs cover W-2 contractors in WA, CA, MA, NY—but systems assume contractors = ineligible
  • Leave coordination is chaos: no central tracking, lost return dates, unclear communication trails
  • Mixed classifications: workers with both W-2 and 1099 assignments—only W-2 hours count for eligibility
  • Manual hour tracking: spreadsheet hell trying to calculate rolling 12-month totals

The compliance risk is real. The operational pain is constant. The tools? They don't exist.

The Solution

ContingentAI is a full-stack workforce analytics prototype that solves what everyone else ignores:

  • Worker management: contractors, FTEs, EOR placements—all in one system
  • Vendor performance: track fill rates, quality scores, compliance, and spend
  • Scenario planning: model headcount changes and cost impacts
  • Leave tracking (newest): FMLA/PFML eligibility, return coordination, and communication logs

The goal: give People Ops and Finance teams the visibility they've never had over their contingent workforce—especially leave compliance.

📈 Real-World Impact

Built from operational experience managing:

  • $80M+ in annual contractor payroll (Upwork)
  • 1,000+ contingent workers across 20+ countries
  • Multiple EOR providers with complex classification rules
  • Post-acquisition integration coordinating systems for 100+ workers (Velocity Global)
  • Leave coordination pain at The Mom Project—now solved

Key Metrics

6
New database models for leave tracking
5
State policies (Federal + WA/CA/MA/NY)
Rule-based
ligibility calculations count W-2 hours only (excludes 1099)
7-day
Configurable return reminders (e.g., 7 days out)

Core Features

Built to solve real operations pain for People Ops, Payroll, and Finance teams.

📊 Worker Management

Track contractors, FTEs, and EOR placements in one system.

  • Multi-classification support (W-2 and 1099 per worker)
  • Department and vendor tracking
  • Rate and salary history
  • Status management (active, terminated, on leave)
  • Bulk import and export

🤝 Vendor Performance

Score and compare your staffing agencies and EOR providers.

  • Fill rates and time-to-fill metrics
  • Quality and compliance scoring
  • Spend analytics (YTD totals, markup %)
  • Contract and payment terms tracking
  • Performance comparison dashboards

📈 Scenario Planning

Model workforce changes before you commit.

  • Headcount adjustment modeling (contractors, FTEs, EOR)
  • Cost impact calculations (monthly & annual)
  • Quick scenarios (10%, 30%, 50% growth)
  • Contractor vs. FTE mix analysis
  • Save and compare scenarios

🗓️ Leave Tracking (Newest)

Finally—leave compliance tracking for W-2 contractors.

  • Automatic eligibility tracking (FMLA: 1,250 hours in 12 months)
  • Multi-state PFML support (WA, CA, MA, NY + Federal)
  • W-2 hour aggregation (filters out 1099 hours)
  • Return coordination (expected date, confirmation, system reactivation)
  • Communication audit trail (who was notified, when, about what)
  • Eligibility alerts (workers approaching 1,250-hour threshold)
  • 7-day return warnings (know who's coming back soon)

Leave Tracking: The Feature That Changes Everything

This is the compliance gap no one talks about—and the feature that makes ContingentAI unique.

Why This Matters

Most workforce platforms assume:

  • Contractors = 1099 = ineligible for leave. Wrong. Many are W-2.
  • Contractors = short-term. Wrong. Many work 12+ months.
  • Leave tracking = HR/benefits problem. Wrong. It's a compliance risk.

The reality? Companies use EOR providers to hire W-2 contractors for extended periods. These workers qualify for FMLA and state PFML programs—but nobody's tracking it.

Until now.

What Makes It Different

  • Mixed classification support: Workers can have both W-2 and 1099 assignments—system only counts W-2 hours for eligibility
  • Rolling 12-month calculations: Auto-updates as hours are logged
  • State-specific rules: Different thresholds and benefits for WA, CA, MA, NY
  • Return workflow: Expected date → Confirmation → System reactivation → Returned
  • Communication logs: Track who you notified (client, vendor, payroll) and when
  • Proactive alerts: Know when workers approach 1,250 hours before they request leave

What I Learned

  • Build for real pain: Every feature came from operational frustration at Upwork, Velocity Global, or The Mom Project
  • Database design is critical: Separating worker classifications from leave balances was the right call—it supports complex scenarios cleanly
  • State compliance is messy: Supporting 5 different leave policies (Federal + 4 states) was hard but necessary
  • Audit trails matter: Communication logs solve a real ops problem—"who did I tell about this leave?"
  • Return coordination is underrated: Expected vs. actual return dates + confirmation workflows are the difference between chaos and control
  • Bootstrap makes it real: Polished UI matters for portfolio credibility—this looks like real enterprise software

What's Next

  • Automated recalculations: Nightly jobs to update eligibility as hours change
  • Email notifications: Auto-send return reminders and eligibility alerts
  • Calendar integration: Export leave dates to Google Calendar or Outlook
  • Bulk actions: Approve or process multiple leaves at once
  • Reports and analytics: Leave usage by department, average duration, compliance trends
  • Mobile optimization: Responsive design for tablets and phones
  • More states: Expand beyond WA/CA/MA/NY to all US states with PFML programs

Want the Full Story?

I wrote a deep-dive blog post about building ContingentAI, including the technical decisions, database design, and why this compliance gap exists in the first place.

📖 Read the Story

Let's Talk Workforce Tech

Building HR tech, compliance tools, or workforce analytics? I'd love to talk about the messy reality of contingent workforce management and what actually works in operations.

📧 Email: amybray0315@gmail.com
💼 LinkedIn: linkedin.com/in/amybrayphr